Control structure determination and enhanced self-tuning using the `MUSE´ AI toolkit

The basis of a comprehensive real-time knowledge based system for control structure design and supervisory control is developed using the MUSE AT toolkit. Representation of control engineering knowledge as procedural code in an object-based structure ensures that the knowledge base is easily understood and modified by domain experts. The simple learning capability that has been implemented has proved to be extremely useful and is being developed further as part of continuing research in this area. It is also hoped to extend the capabilities of the supervisory modules particularly in dealing with loop interactions, the prediction and diagnosis of faults and the provision of an intelligent operator interface